Seeds of cucumber (C. sativus) inbred line 3461 (China Agricultural University) were used in the experiments. Cucumber seeds were surface sterilized and sown in a autoclaved soil (50 nutritive soil and 50 % vermiculite) at 26 ± 2°C in a greenhouse. Single cucumber seedlings with two leaves were transplanted into a 15-cm diameter plastic pot filled with steam-sterilized river sand and randomly placed in a greenhouse, provided with a regular regime of 16 h of light/day and temperature maintained between 26 and 30°C. Plants were regularly watered with Hoagland's solution. For the cytochalasin D treatments, a previously reported method was used (de Almeida Engler et al., 2004), with minor modifications. Individual plants were pre-treated on medium with 2 μM cytochalasin D (Sigma, Beijing, China) for 1 day before being transplanted into 15-cm diameter plastic pots, after that 50 ml 0.5 μM cytochalasin D was added every day at 9 am.

An M. incognita race 3 from Institute of Plant Protection, China Academy of Agricultural Sciences population had been reared previously in a glasshouse on susceptible cucumber plants (Liu et al., 2015). The progression of nematode infection is shown in Supplementary Figure 1. Nematode inoculation and determination of infestation levels were as previously described (Bybd et al., 1983; Molinari et al., 2014). Nematodes were distinguished as motile vermiform individuals (J2s), swollen individuals that had become sedentary (third and fourth stages, SJs) and adult females (AFs) (Molinari et al., 2014).

Paraffin embedding technique was used to achieved the paraffin section of roots and galls. Cucumber roots and galls at different time were fixed, embedded, and sectioned as previously described (Kerk et al., 2003; Zhang et al., 2013).

Cucumber roots and galls were fixed, embedded, and sectioned as previously described (Liu et al., 2016). Thin cross-sections were made with a UC6I microtome (Leica, Germany) and examined with a JEM-123O scanning transmission electron microscope (JEOL, Japan).

Sequence information for 11 A. thaliana ADF genes were retrieved from previous study (Ruzicka et al., 2007). To assess the corresponding cucumber orthologs, a basic local alignment search tool (BLAST) was performed for each one of the A.thaliana ADF protein sequences against the Cucumber Genome Database (http://cucumber.genomics.org.cn/page/cucumber/index.jsp). Those with the highest amino acid percentage identity were selected. Cucumber ADFs were named according to the annotation of the Cucumber Genome Database. The multiple sequence alignment of CsGID1s and related CsGID1 proteins was performed using ClustalW within the MEGA5 software package (Tamura et al., 2011), and the highlighting conserved sequences were drawn by using BoxShade (http://www.ch.embnet.org/software/BOX_form.html). The phylogenetic tree was constructed using the Neighbor-Joining (NJ) method (Saitou and Nei, 1987) with the Poisson model and 1000 bootstrap replicates test by MEGA5 (Liu et al., 2016). The gene numbers corresponding to every CsADF gene in the Cucumber Genome Database are shown in Supplementary Table 1. The PCR primers used for cloning CsADF genes are listed in Supplementary Table 2.

The fusion 35S::CsADF2-3::GFP, 35S:: CsADF7-1::GFP, 35S::CsADF6::GFP, 35S::CsADF5:GFP constructs and the control plasmid of 35S::GFP were transformed into onion epidermal cells as previously described (Zhang et al., 2014a). The PCR primers used for cloning CsADF genes are listed in Supplementary Table 2.

Total RNA was extracted from roots or galls using the SV Total RNA Isolation System (Promega, Beijing, China), and cDNA was synthesized using the MultiScribe™ reverse transcriptase kit (Applied Biosystems, Beijing, China). Quantitative real-time RT-PCR (qRT-PCR) was performed as previously described (Liu et al., 2016), The gene specific primers used are listed in Supplementary Table 2.

Using A.thaliana ADFs protein sequences, we identified and named eight candidates ADF genes based on a BLAST search of the Cucumber Genome Database (http://cucumber.genomics.org.cn/page/cucumber/index.jsp). All ADF genes from the cucumber were anchored to cucumber chromosomes, the chromosomal locations and the scaffold sequence directions of the 8 CsADF genes were analyzed using BLASTN. Further analysis of the cucumber ADF genes revealed that they were distributed on five of the seven cucumber chromosomes and none of the CsADF gene was found on chromosomes 6 and 7 (Figure 1A). We observed homologous CsADF genes located on different chromosomes in cucumber, suggesting that duplicated events were potentially involved in the evolution of cucumber CsADFs.

Previous studies have shown that ADF genes contain a conserved region called the H-domain (Lappalainen et al., 1998), In cucumber, it is present in all ADF genes (Figure 1B). As was reported to be the case with A. thaliana ADF orthologous genes, all CsADFs was predicted to contain three exons and two introns, with the exception of CsADF2-1 (having seven exons and six introns) and CsADF2-2 (having four exons and three introns; Figure 1C). Subsequently we further analyzed the phylogenetic relationship with A. thaliana. The results revealed that the cucumber ADF gene family is grouped phylogenetically into four ancient subclasses supported by the existence of homologous ADF protein sequences from each subclass in the A. thaliana. (Figure 1D). For example, CsADF6 are significantly more related in protein sequence to Arabidopsis ADF6 in subclass III than other cucumber ADF genes.

To further determine the subcellular localization of CsADF proteins, sequences corresponding to 35S::CsADF2-3::GFP (subclass I), 35S::CsADF7-1::GFP (subclass II), 35S::CsADF6:GFP (subclass III) and 35S::CsADF5:GFP (subclass IV) fusion proteins were constructed and transiently expressed in onion epidermal using particle bombardment (Zhang et al., 2014a). Transformation with the 35S:GFP control construct resulted in GFP localization throughout the cells (Figures 2A–C), while green fluorescence resulting from expression of each of the four ADF-GFP fusion proteins was localized in both the cytoplasm and the nucleus, with a characteristic filamentous appearance (Figures 2D–O), consistent with a previous report (Tang et al., 2016).

Transcript expression analysis by a quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) experiment revealed that all CsADF genes are expressed in roots. Among them, CsADF1, CsADF2-1, CsADF2-2, CsADF2-3 (Subclass I genes), and CsADF6 (Subclass III gene) have a higher expression, while CsADF7-1, CsADF7-2 (Subclass II genes), and CsADF5 (Subclass IV gene) have a lower transcript (Figure 3A). To investigate whether M. incognita infection correlates with changes in the transcription of ADF genes, we quantified the expression of eight ADF isotypes using qRT-PCR in roots and galls 7, 14, and 21 days after inoculation with 300 M. incognita active J2s (DAI; Figure 3B). The transcript levels of five isotypes (CsADF1, CsADF2-2, CsADF2-3, CsADF5, and CsADF6) were higher in galls of the infected roots compared with those of the control. Members of subclass I genes (CsADF1, CsADF2-1, CsADF2-2, and CsADF2-3) except CsADF2-1 exhibited a specific induction at 14 DAI. The expression of subclass II genes (CsADF7-1 and CsADF7-2) showed no significant change. CsADF6 (Subclass III gene) showed a specific induction at 21 DAI. It's interesting that CsADF5 (Subclass IV gene) was weakly expressed in root, but its transcript level was remarkably up-regulated as early as 7 DAI.

Previous study showed that microtubules and actin filaments appeared partially depolymerized throughout the feeding site development (de Almeida Engler et al., 2004), and our results indicated that nematode infection induced the expression of five CsADF genes (Figure 3B). Given that cytoskeletal-inhibitor treatments result in the cessation of the development of normal giant cells in Arabidopsis (de Almeida Engler et al., 2004), we hypothesized that cytoskeleton inhibitors are likely to reduce the expression of CsADF genes. This idea was supported by the qRT-PCR experiment revealing that an approximately 2-fold down-regulation of the CsADF genes occurred in roots of treated plants (Figure 3C).

As shown in Figure 4A, infection with M. incognita resulted in cucumber root galls; however, in cytochalasin D-treated plantlets, the galls were smaller than in untreated controls. A detailed morphological comparison was performed of nematode infected roots and controls (Figures 4B–I). Cells in both longitudinal and cross sections of control roots were regularly arranged (Figures 4B,D), but in nematode infected root galls, most of the cells were less organized and giant cells were induced at feeding sites (Figures 4C,E). To provide higher resolution of cellular structures, root cells were examined by transmission electron microscopy (TEM) (Figures 4F,G). We focused on variation in the appearance of the cell walls and plasma membrane as several studies have shown that their morphology is particularly important for understanding giant cell development (Jones and Dropkin, 1975, 1976; Jones and Payne, 1978). In the control roots, each cell had a distinct and normal cell wall (Figure 4F), while in the infected roots, ingrowths and degradation occurred on the walls of the giant cells, such that the wall itself was sparse (Figure 4G). The numerous ingrowths greatly increased the surface area of the plasma membrane and were most pronounced adjacent to the vascular tissue, which is a potential site of nutrient influx into giant cells (Baldwin, 1978). Treatment of uninfected roots with cytochalasin D caused cellular defects (Figure 4D), such as a wider root central cylinder and larger cell files. Giant cells in roots at the same stage of infected seedling treated with cytochalasin D were much smaller than those of uninfected and treated roots (Figure 4I) and there were fewer of them than in untreated plants (Figure 4E).

It has been reported that cytochalasin D treatment blocks cytoskeleton dynamics, resulting in the cessation of development of normal giant cell in Arabidopsis (de Almeida Engler et al., 2004). We investigated whether cytochalasin D treatment similarly affected RKN parasitism during a compatible interaction with cucumber. Individual plants were pre-treated on media with high concentrations of cytochalasin D (2 μM) for 1 day, after which they were transferred to soil and inoculated with 300 active J2s. Next, 50 ml 0.5 μM cytochalasin D was added every day. Seven and 21 days after inoculation, we counted the number of nematodes that had penetrated the roots (Figures 4J,K). At day 7 after inoculation, there was no significant difference in total nematode number and the numbers of motile invasive forms (J2s) between untreated and cytochalasin D-treated plants (Figure 4J). We observed very few invasive juveniles as they had become sedentary in the roots of untreated plants. However, even fewer sedentary juveniles (SJs) were present in the cytochalasin D treated plants (Figure 4J). Twenty-one days after inoculation, the number of juveniles developing in the roots increased substantially, whereas the presence of motile invasive forms was reduced to a small fraction. However, the roots of cytochalasin D treated plants exhibited a significant reduction in the number of both J2s and SJs compared with untreated plants (Figure 4K). Although very few AFs were present at this stage, no differences were evident between treated and untreated plants. In general, the number of juveniles that penetrated the roots 21 days after inoculation was approximately two-thirds of the number of inoculated juveniles in the untreated plants and one-half of the number of inoculated juveniles in the treated plants.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.